Background
This is a wreck that Tyler discovered off Point Loma about three (?) years ago so it isn’t anything new. We went to dive it after the dive on the unidentified object. I had never been on this wreck and figured it might make sense to build a photogrammetry model for others to use as a reference if they dive it.
It is about 20 feet wide and maybe 100 feet long. The water color was “green” for our dive which shows through in the model and photos.
The other characteristic which became obvious only in the context of seeing the overall model is that the barge is not flat. It is convex with the center being higher than the edges. I’m guessing that this might be to help carry heavier loads and not “bow” under the load. This is similar to some “flat bed” trailers which are actually convex and not flat.
The square openings are also interesting. I’m not sure if they are some kind of drain holes or what function they serve. I’m certainly not an expert on barges so maybe one of the readers will have some more information. If you look closely, there are also some small slits around the larger square openings.
Below are some photos, ccreencaptures of the model, and a link to the on-line model.
Model and Photos
The “live” model is posted online at Sketchfab at the link below:
Model Screenshots
Here are some screenshots from the model. You can clearly see that it is skewed somewhat and isn’t a great model. Normally, I would “fix” this but since it is only a barge, the quality is good enough for people to get a sense of the structure.
Photos
A couple quick photos taken while shooting for the model. You can see how “green” the water is. The first photo is straight out of the camera so you can see the “actual” color.
Here are some additional photos after a little work in Lightroom (not much though):
Detailed Build Process
This is probably only interesting to those with a desire to learn more about the process of photogrammetry and some of the challenges.
It was a bit challenging to build a model on this structure in a single dive. I wasn’t trying to make a perfect model and sometimes large, flat surfaces can be hard to keep track of what is covered and what is not. Also, long stretches of space lend themselves to slight misalignments which can then veer off course.
This happened in the first model that I built which was using “Medium” accuracy. All of the photos “aligned” in Metashape but clearly there was a slight skew. Below is the sparse cloud:
Version 2 of the model got even weirder when I used High accuracy (which doesn’t really make sense to me). This version did NOT use the Metashape option for “Generic preselection.”
I then built a model which enabled “Generic preselection” and got a version that was good enough for the purpose of documenting the wreck. There is still a slight misalignment, but not too bad.
Here was the sparse cloud and the finished textured mesh:
It isn’t perfect, but it is good enough to give readers a sense of the structure.
Conclusion
Will I go back? Nope, not a chance.
It was okay to document the wreck, but not worth another dive for me. It would make a good dive site for advanced open water training since it is large, easy to find, navigate, etc.